The system now supports global R scripts, to be run after the successful connection to R. Each connection will have all listed R scripts run. Per-user R scripts are also supported. This allows scripts specific for a user to be run after that user logs in. Look in the example configuration file for more information.
It appears that the readFile function requires a second parameter for it to appear as text in the result.
R v2.10 removed the pre-generated HTML help as a standard component of R, and instead moved the HTML help to a standalone server. This patch implements support for this help server, once it is started up as an entirely separate R instance (i.e. not from R-Node). I tried both starting up the R http help server from R via Rserve, and also via a normal spawned R instance, but in both cases the R http server would never respond to any requests. So the R help server now needs to be a separate R instance, by default running on port 22900.
Basically make a whole bunch of small changes that allow most of the R-Intro book to run.
expressions and other things, like boolean expressions.
We're just gonna have to find another way to deal with code that needs the raw objects.
They are now shown as part of the normal graph list, and can be downloaded. Default sizing isn't quite right though.
are provided in. This also works to allow users to graph all manner of graphs server side and have them show up in the client. The user can also turn off client side graphs (for those graphs that exist client side), and only use server side graphs. Finally, a stupid bug was fixed which had all the R command handlers deriving from the wrong superclass.
commands. No longer needed.
This is done so that we can allow the server to deal with a request in a more logical fashion if deems it reasonable, e.g. for graphical commands.
the standard R console more closely - if we don't understand how to deal with a command, then we print out the results.
If we recognise the request as a request to graph something, then we'll wrap it in code that'll generate a PNG, which we then return to the user. More to do on this (to support multiple graphing commands that should all generate a single graph).